日本語版

Topic hub — 2026-09-19

Is scorched rice bitter?

The layer of rice stuck to the bottom of the pot has its own name in many places: guoba in China, nurungji in Korea, okoge in Japan, socarrat in Valencia. Every one of those words comes from a verb meaning to scorch, and the dictionaries keep the scorching in the definition.

And yet almost nothing in the material gathered here speaks of bitterness. What it speaks of is fragrance. In the one case where a trained panel evaluated it, bitterness was not on the scale at all. This page does not try to fill that gap; it lays the material out so the gap is visible. The word bitter appears almost only in talk of scorching something too far.

What this page establishes

• The dictionary definitions of guoba and socarrat both use the word for scorching and contain no taste term. The Valencian Academy of Culture defines it positively: partially burnt, yet at times very tasty.
• In the one case where 9 trained panellists evaluated socarrat, the roasted attribute rose in intensity while the burnt attribute stayed below the white-rice control. Bitterness was not among the attributes.
• The consumer acceptance test for nurungji (n=104) likewise placed no bitterness among its 9 attributes. Savoury character is treated as the core attribute.
• Bitterness appears on the over-scorched side. One of the few explicit statements holds that moderate scorching is a treasury of umami and aroma while over-scorching turns bitter.
• On the smoke side of the same heat, bitterness is measured. In hot-smoked mackerel a trained panel barely perceived it, and what did raise it was a change in the salt composition.

The word says scorched, the judgement says fragrant

Start with the definitions. Guoba is defined as the layer of rice that sticks to the bottom of the pot and scorches while the rice steams, usually pale yellow or black. The scorching is named; no taste term is.

Socarrat is the same. The Valencian Language Academy defines it as the layer of paella rice that remains stuck to the vessel it is cooked in, and explains that the diminutive concentrates the flavour of the essence of paella. Here too there is no specific taste term for bitterness or fragrance.

A second normative dictionary goes one step further and approves. Partially burnt, yet at times very tasty. It uses the word burnt and hands out a verdict of good eating.

Food writing makes the distinction more explicit. Of guoba: the character for scorch is there, but this is not the black-charcoal scorch of burning down the kitchen; it is a golden fragrant scorch, an aroma and not a bitterness. This is one writer's account with no sensory panel behind it, but the intent to separate the word scorch from bitterness is unmistakable.

Japanese okoge runs the same way: lightly scorched to the point where the pot bottom turns somewhat brown is fragrant and has flavour, while complete carbonisation is undesirable because the smell gets into all the rice. The contrast drawn is between fragrance and an unpleasant smell, and the word bitter is not used.

What happens at the bottom of the pot

The explanation is consistent across regions. A Korean encyclopedia states that once the pot bottom passes 200 degC, the amino-carbonyl reaction browns the rice while producing volatile carbonyl compounds, and that these soak into the rice to form its characteristic savoury aroma. Bitterness is not mentioned.

The Japanese technical account explains it the same way, as sugars and amino acids combining to produce the brown melanoidins and aroma compounds, and calls the result crisp and fragrant.

In the peer-reviewed work the conditions are pinned to numbers. Across 5 heating conditions of 2, 5, 8, 11 and 14 minutes, the moisture content of nurungji fell sharply from 25.14% at 2 minutes to 0.24%, with sugar content of 1.40 to 3.00 degrees Brix. The browning is described as forming an appetising colour and a savoury flavour, but that too is not a formal sensory evaluation.

One study goes down to the aroma compounds, though not of guoba itself. It names the characteristic aroma of a Chinese post-fermented tea "guoba aroma" and identifies its principal odorants as 5 pyrazines carrying a roasted note. That guoba serves as the reference for an aroma is itself a fact about how this subject is treated.

Where it was measured, bitterness was not on the scale

The heaviest fact on this page is here. In the one case where a trained panel evaluated the crust at the bottom of a pot, bitterness was not part of the scale.

With 9 trained panellists rating on an unstructured 10 cm scale, the roasted attribute rose significantly in intensity through the cooking stages where socarrat forms, while the burnt attribute stayed below the white-rice control. The paper reads this as suggesting that the socarrat forms a crust in the pan that keeps the rice from burning. The attributes evaluated were roasted and burnt; bitterness is not among them.

The consumer test is the same. For 5 nurungji samples blended with 0 to 100% barley, the 9 attributes rated by 104 consumers were overall acceptance, appearance, aroma, taste, sweetness intensity, texture, savoury intensity, aftertaste and purchase intention. Bitterness was not set as an attribute. It was not denied; it was not measured.

From the compound side there is one thread that comes close. The furans, furanones and ketones detected late in the cooking are a group reported in earlier literature to carry a bitter-caramel aroma alongside roasted and caramel notes. But that link comes from cited literature rather than the paper's own panel, and it is not data on anyone tasting socarrat as bitter.

The over-scorched side

Explicit mentions of bitterness are rare in this material. One of them holds that the fragrance of okoge arises from the Maillard and caramelisation reactions, and that over-scorching turns bitter and unpleasant while moderate scorching is a treasury of umami and aroma. The source is a personally run blog whose author and expertise are unknown.

So the line connecting scorching to bitterness is drawn, in this material, only past the point of going too far. What these dishes name is the state short of that line.

On the smoke side it is measured

In the same family of heat, the smoke side does have bitterness ratings. When 6 to 11 trained panellists rated multi-stage hot-smoked Atlantic mackerel across 21 attributes, strong smoky odour and flavour appeared while deterioration attributes such as bitterness and rancidity were barely perceived. The ratings are low but not zero.

So what did move the bitterness? In chub mackerel hot-smoked under the same conditions, bitterness intensity was 0.0 to 0.3 with sodium chloride alone or 25% potassium chloride substitution, and about 2 to 2.6 at 50% and 75% substitution. What moved was the salt composition, not the smoke.

In showing that heat does not necessarily create bitterness, these two point the same way as the pot-bottom material. The crossing called roasted and bitter is, on the cooking side, not yet supported by measurement.

Industrialised scorching has been measured

Bitterness has not been measured in the crust at the bottom of a pot. But in a product that runs the same reaction industrially, the bitterness has been traced to its compounds: caramel colours.

JECFA divides them into 4 classes by method of manufacture. Class I is carbohydrate heated under control with alkali or acid; Class II with sulfite compounds; Class III with ammonium compounds; Class IV with both.

For Class III, a sensory-guided fractionation has been carried out. DDMP, HMF, furfural and 5-methylfurfural were isolated and structurally determined as the main bitter compounds, with threshold concentrations reported at 14.3, 30.0, 4.4 and 17.5 µmol/L respectively. Nothing at this level exists for the crust at the bottom of a pot.

The sensory side has been measured too. Added at 0.20% (w/v) to huangjiu, Classes I, III and IV all raised the bitterness intensity significantly against the untreated control.

More of it means more bitterness. Raising the addition across 0.03%, 0.10% and 0.20% (w/v) gave a significant increase in bitterness intensity with dose in each class.

So the bitterness of scorching is not beyond measurement. That it has not been measured in the pot-bottom crust reads less as an impossibility than as a question nobody set out to ask on that side.

Questions people ask

Is okoge bitter?

Within the material gathered here there is no record of anyone measuring it as bitter. The dictionary definitions carry no taste term, the food writing speaks of fragrance, and in the one trained-panel evaluation bitterness was not on the scale. Where the word appears, it appears as over-scorching turning bitter, and that source is a personally run blog.

How do fragrance and bitterness differ here?

In this material, fragrance is used as a word for aroma and bitterness as a word for taste. The socarrat panel rated roasted and burnt, which are aroma attributes, not taste attributes. When guoba is described as an aroma and not a bitterness, that is the distinction being drawn.

Does more scorching mean more bitterness?

It has not been measured. The nurungji experiment shows moisture falling from 25.14% to 0.24% across 5 conditions from 2 to 14 minutes, but bitterness was not evaluated. In the socarrat study the roasted attribute rose with cooking while the burnt attribute stayed below the white-rice control. The statement that going too far turns it bitter sits on the side of talk, not measurement.

Is smoked food the same story?

The smoke side has measurements. In hot-smoked mackerel, deterioration attributes including bitterness were barely perceived. Bitterness rose under the same hot-smoking conditions when potassium chloride substitution reached 50% and 75%, moving from 0.0 to 0.3 up to about 2 to 2.6. Neither smoke nor scorch moved it; salt did.

What is not known

• Not one piece of data in this material has a human rating the crust at the bottom of a pot as bitter. It is not that bitterness is absent; it has not been measured.
• The link between the furans, furanones and ketones found in socarrat and bitterness comes from cited literature, and whether those compounds are perceived as bitter within socarrat has not been established.
• Where the boundary of over-scorching lies is unknown. No source here gives it in temperature, in time or in colour.
• Whether guoba, nurungji, okoge and socarrat are regional names for one phenomenon or different things made under different heating conditions cannot be decided, because nothing here compares them.
• The low bitterness ratings in hot-smoked mackerel are not a comparison against an untreated control. What smoking itself does to bitterness cannot be said from this material.
• Whether the bitter compounds identified in caramel colours arise in the same way in a pot-bottom crust has not been established. The huangjiu addition test is a different condition from eating scorched rice.

Read in the encyclopedia

• Okoge — the Japanese pot-bottom crust, with a line drawn between fragrance and carbonisation
• Guoba — the Chinese crust, used even as the reference aroma in tea evaluation
• Nurungji — the Korean crust, the only entry with heating conditions and consumer-test figures
• Socarrat — the paella pan bottom, the only case evaluated by a trained panel
• Hot smoking — the smoke side, the process where bitterness is actually rated
• Caramel colours — industrialised scorching, traced to bitter compounds and their thresholds

Go deeper in the research notes

• The boundary between roast aroma and bitterness — separating what is judged as aroma from what is judged as taste
• Does roasting make bitterness? — comparing the direction of heat and bitterness in cacao, coffee and almonds
• Can cooking create bitterness? — placing scorching in the four quadrants of adding, removing, designing and suppressing
• Butter roasting and kopi — from the side of techniques that stop just short of scorching

Source ledger

The primary sources and limitations for the claims on this page can be checked from the source links inside each card. The full list is on the Japanesesource ledger.